Google noodle shop is open! Uncover the machine learning technology behind Ramen

Time:2021-10-21

Google noodle shop is open! Uncover the machine learning technology behind Ramen

More code, more vitality!
Google noodle shop officially “opened”

Google noodle shop is open! Uncover the machine learning technology behind Ramen

Want to experience the pleasure of a ramen master?
Now go  Official website of 2020 Google Developer Conference
Create your own “code Ramen” now!

Google noodle shop is open! Uncover the machine learning technology behind Ramen

Come to Google noodle shop to experience virtual Ramen

Combining work and rest makes you healthier physically and mentally

continuitySix Day Google Developer ConferenceIt not only brings full technical dry goods, but also encourages developers who fight with the code day and night to combine work and rest. They can also relax and replenish their vitality during the interval of watching the speech.Posenet in tensorflow.js is used to realize dynamic capture, players only need to participate in the interaction from webcam and browser to easily experience the happy “code Ramen”.

Google noodle shop is open! Uncover the machine learning technology behind Ramen
Google noodle shop is open! Uncover the machine learning technology behind Ramen
Google noodle shop is open! Uncover the machine learning technology behind Ramen
Google noodle shop is open! Uncover the machine learning technology behind Ramen

How do you pull noodles when you come to Google noodle shop? Each round of ramen is limited to 60 seconds. Players can imagine that they are Ramen masters. They can make virtual “noodles” through the action of moving hands and upper body, such as as as much as possibleMove up and down to pull the noodlesYes, you canAccelerate the swing to perform cool Ramen stunt

Google noodle shop is open! Uncover the machine learning technology behind Ramen

In the process of noodle drawing, you should always cooperate with your hands to collect special ingredients to obtain points to increase the length of noodles, but be careful that if you receive “dark cooking”, the length of noodles will be shortened. The higher the score, the more you can unlock“Ramen hero badge”! Can’t wait to “decide the length” of Google noodle shop? Using the computer to participate in the game experience is better!

Tensorflow.js realizes dynamic capture

Create “code Ramen”

Google noodle shop isDynamic capture is realized based on posenet in tensorflow.js, combine the user’s body movements with the concept of “Ramen” to create an interesting and fun interactive experience.

Google noodle shop is open! Uncover the machine learning technology behind Ramen
Google noodle shop is open! Uncover the machine learning technology behind Ramen

The key to the whole experience isDetection of hand movement。 By selecting the rehearsed model from posenet and estimating the hand position in real time on the device camera, tensorflow.js can run the existing machine learning model in the browser and realize hand motion detection without spending too much energy. Through this series of technical support, players can move, stretch and twist virtual noodles in the game.

Google noodle shop is open! Uncover the machine learning technology behind Ramen

It is also used in Google noodle shop  Angular constructs the main framework, set up a powerful infrastructure for website functions, and thenCombine tensorflow.js with the rich model library in posenetFrom motion detection gestures and various limb movements, pose directly from webcam and browser to create real-time visual modeling of human beings. And game rendering and physics engines (such as phaser) are also used to help realize the real physical state of noodles one by one in the virtual environment.

Angular 
A modern JavaScript framework for building dynamic and interactive applications in HTML. It can use the web page as a template language, expand the syntax of HTML, and express the components of the application clearly and concisely.

TensorFlow.js 
An open source library for machine learning using JavaScript and advanced APIs. If you are not familiar with machine learning technology, this is a great beginner tool to define, run and train models according to your needs.

Google noodle shop is open! Uncover the machine learning technology behind Ramen

PoseNet

A visual model for detecting and predicting single or multiple person gestures in an image or video by marking key body positions. The technology can not recognize the characters in the image. The algorithm only detects the key points indexed by “part ID”, and the confidence score is between 0.0 and 1.0, of which 1.0 is the highest.Click to learn how motion capture can bring SVG characters to life in the browser!

Google noodle shop is open! Uncover the machine learning technology behind Ramen

The birth of Google noodle shop is just the beginning. Google expects that through the interactive experience of Google noodle shop, more developers will be inspired to work with tensorflow open source code community to create more unimaginable possibilities with technology and explore more extraordinary interactive experience. Not much to say,Click 2020 Google together Official website of Developer ConferencePull up the noodles!

2020 Google Official website address of developer conference:https://developersummit.googlecnapps.cn

Google noodle shop is open! Uncover the machine learning technology behind Ramen